Floral hosts of leaf-cutter bees (Megachilidae) in a biodiversity hotspot revealed by pollen DNA metabarcoding of historic specimens

The DNA extraction protocol was optimized in a previous study [26] and used here to ensure even rupture of all pollen grains across taxa. The DNeasy1 Plant Mini Kit (Qiagen, Hil- den, Germany) was used without bead disruption during lysis. DNA was eluted in 50 μl of buffer EB, with reapplication of the eluate to the DNeasy Mini Spin Column for a second elu- tion step to increase DNA yield. The nuclear ITS2 region was selected as the barcode to be targeted for the identification of pollen’s plant origins. The primers for ITS2 [27] were modified to include overhang adapters as described in the Illumina 16S Metagenomic Sequencing Library Preparation Guide (Illumina, 2013). These overhang adapters allow the primers to be used directly in the standard Illumina indexing and adapter PCR. The primers used were ITS3F_Tag_IL 5’ TCG TCG GCA GCG TCA GAT GTG TAT AAG AGA CAG GCA TCG ATG AAG AAC GCA GC 3’ and ITS4R_Tag_IL 5’ GTC TCG TGG GCT CGG AGA TGT GTA TAA GAG ACA GTC CTC CGC TTA TTG ATA TGC 3’ (overhang adapters indicated in bold and underlined). Oligonucleotides were synthesised by Inqaba Biotechnical Industries (Pty) Ltd (Pretoria, South Africa). Barcode amplification was achieved in reactions with a final concentration of 1 × Phusion1 High Fidelity PCR Master Mix with HF Buffer (Thermo Scientific, Waltham, MA, USA), 0.5 μM of each primer, and 5 μl of DNA template. Milli- Q1 H2O (Merck Millipore, KGaA, Darmstadt, Germany) was added to a final reaction volume of 50 μl. PCR conditions were as follows: 98˚C for 3 minutes, followed by 30 cycles of 98˚C for 7s, 65˚C for 30s and 72˚C for

30s, with a final step of 72˚C for 10 min. Negative controls were included in PCR-based steps. Amplified barcodes were visualised using 2% agarose gel electrophoresis. Amplification products were purified using the Agencourt AMPure XP (Beckman Coulter, Brea, California, USA) bead purification system according to the manufacturer’s protocol. The DNA concentration of approximately half of the samples was evaluated using a Qubit1 2.0 Fluorometer (Invitrogen, Life Technologies, Carlsbad, CA, USA) and a Qubit1 dsDNA High Sensitivity Assay Kit (Invitrogen, Life Technologies). Samples were randomly selected for eval- uation and both kits were used according to manufacturer instructions. Pollen from five of the six bee species was sequenced using an Illumina HiSeq 2500. Pollen from M. venusta samples were sequenced using an Illumina MiSeq and classified in a previ- ous study. Nextera XT (Illumina, Inc. San Diego, CA, USA) indexes were added according to the sequencing preparation protocol (Illumina, 2013) to multiplex samples. Indexed samples were pooled equimolarly and sequenced on a HiSeq 2500 sequencer (Illumina, Inc.) at the ARC’s Biotechnology Platform, Pretoria, South Africa, using the HiSeq Rapid SBS Kit v2, with 2 × 250 bp paired end reads (Illumina, Inc.). Bioinformatics and statistical analyses Demultiplexing of samples was done with CASAVA v1.8.2 (Illumina Part #15011196 Rev D) based on the Nextera index sequences used. Low quality bases and adapter sequences were trimmed from reads with Trimmomatic 0.35 [28] using a sliding window of eight bases with

an average quality of 20 bases required per window. Trimmed reads were merged in Mac- Qiime 1.9.1–20150604 [29] using the multiple_join_paired_ends.py script. Trimmed sequence data is available from FigShare (see Data Availability Section). The ITS2 database for Viridiplantae [3] was used for identification of pollen samples. Clas- sification of sequences was performed using the Ribosomal Database Project (RDP) classifier v. 2.10.1 [30]. Analysis was performed following the bioinformatics workflow described by [30]. Reads not meeting the required 0.8 confidence level at family level after classification, and rare taxa (less than 0.1% of the total amount of reads identified as plant) were removed PLOS ONE Floral hosts of Megachilidae revealed by metabarcoding PLOS ONE | https://doi.org/10.1371/journal.pone.0244973 January 21, 2021 4 / 16